A nine unit gets appraised two different ways depending on who's looking

Something I ran into on a 9-unit in a secondary market that I haven't resolved.

The guide logic is that at five units you're valued on NOI and a cap rate. That's how I underwrote it. In-place NOI 84k, market cap in that submarket around 6.25, so 1.34M. The appraisal came back at 1.14M, and when I read the reconciliation the appraiser had weighted sales comparison at 60 percent and income at 40 percent, because there were only two income-verified multifamily sales in the county in the relevant window and both were smaller. The sales comps put it at about 122k a unit. Per-unit pricing on a 9-unit in a market with almost no trades is closer to a residential comp exercise than an income exercise.

What that means for my model is that forcing NOI does not reliably convert to value on schedule. I can raise NOI 15k and add 240k of theoretical value at a 6.25 cap, and if the appraiser is still leaning on a per-unit sales grid, my refi proceeds move by a fraction of that. The whole value-add lever in this tier assumes the income approach is what governs, and in thin markets it partly doesn't.

What I'm weighing now is whether to only underwrite value-add in markets deep enough to have real income comps, which pushes me toward larger properties and more competition, or to accept a wider gap between NOI value and appraised value and size my equity for it. I don't have a clean answer. Curious whether people are seeing the same weighting in markets they buy in, and whether anything in the appraisal engagement can push it toward income, like providing rent rolls on comparable properties yourself.